Hubspot, Inc.

Amazon Technologies, Inc.

Jake Kramer

Kenneth John Basye

Alexander David Rosen

Michael Axiak

Tom Monaghan

Paul Weston

Xi Wang

Mark Greene

Provided herein is an electronic message management platform that enables calculating a size of a sample of a message recipient list that delivers a confidence level greater than threshold for the entire message recipient list by applying a predicted bounce percent for the list to a sample size selection algorithm, selecting a total of recipients in the message recipient list that is consistent with the sample size, processing messages to each of the recipients in the sample, monitoring message life cycle events for the recipients in the sample set over a period of time, and based on the monitoring taking an action with the message recipient list selected from a list of recipient list actions consisting of processing remaining recipients in the message recipient list and halting processing of the message recipient list.

Mitigating abuse in an electronic message delivery environment

Provided herein is a method of preparing a route for a new message through a plurality of networked nodes in an Internet Protocol-based network including calculating a recipient quality score based on a message engagement event with a past message, applying at least one model to determine a probability of the recipient engaging with the new message, based on at least one of the recipient quality score or the determined probability, determining a quality metric, and selecting the route from among a plurality of quality-specific tranches of available routes based on a prediction of the probability of engagement for the new message by applying a machine learning system to determine which route produces a greater likelihood of engagement for parameters of the new message than other routes. New messages for recipients with a first quality metric are sent using a different route than recipients with a second quality metric.

Quality-based routing of electronic messages

In a dynamic automatic speech recognition (ASR) processing system, ASR processing may be configured to estimate a latency of returning speech results to a user based on work being done by an ASR processor. The ASR processing system may measure work done by an ASR processor by measuring one or more time independent metrics and comparing the metrics to threshold values. If the metrics exceed the thresholds, the ASR system may take steps to reduce latency associated with processing the utterance, including adjusting a speech recognition parameter.

Dynamic pruning in speech recognition

Growing community of inventors

Cambridge, MA, United States of America

Jake Kramer